github.com/kubernetes-retired/kubefed
v0.0.10
#780 most downloaded on Go modules
kubernetes-retired/kubefed
What this package is like to depend on
Last release 7 years ago
no release in 18 months
Ships fairly regularly
a new release about every 3 weeks
Some releases are documented
notes for 3 of 10 stable releases
Nothing withdrawn
no release was ever pulled
8 years old
20 releases · first in 2018
0 releases in the last 12 months
see the full history below
Release timeline
20 releases · Jul 2018 to Aug 2019Releases
latest 20-
v0.1.0-rc617 Aug 2019 pre-releaseRelease notes
Open source →- #1099 Updates to propagation status are now only made in response to propagation to member clusters or errors in propagation. Previously propagation status was updated every time a federated resource was reconciled which could result in unnecessary resource consumption.
- #1098 Propagated version is now only updated when changed.
- #1087
The ReplicaScheduling controller now correctly updates existing
overrides of
/spec/replicas. Previously the controller was able to create and remove overrides for thereplicasfield but would fail to update them. - #1076
All
kubefedctlcommands now default--host-cluster-contextto the current context in log messages. - #1086
kubefedctl federatenow removes all metadata fields except labels from the template of federated resources created from a non-federated resource. Previouslymetadata.annotationsandmetadata.finalizerswere not removed which could result in propagation errors. - #1079 The spec field is now required in generated federated types. For types generated previously, a check has been added so that a missing spec field does not cause a nil pointer exception.
-
v0.1.0-rc501 Aug 2019 pre-releaseRelease notes
Open source →- #1058 KubeFedConfig spec.scope is now immutable.
- #1052
Support has been added for varying the
apiVersionof target resources. This is intended to allow a federated type to manage more than one version of the target type across member clusters.apiVersioncan be set either in the template of a federated resource or via override. - #951 Propagation status for a namespaced federated resource whose containing namespace is not federated now indicates an unhealthy state.
- #1053 API group changed from kubefed.k8s.io to kubefed.io.
-
v0.1.0-rc419 Jul 2019 pre-releaseRelease notes
Open source →- #908 Adds admission webhook validations for KubeFedCluster API.
- #982 To ensure compatibility with controllers in member clusters, metadata.finalizers and metadata.annotations can no longer be set from the template of a federated resource and values for these fields are always retained. The addition of jsonpatch overrides ensures that it is still possible to add or remove entries from these collections.
- #1013 Add support for defaulting KubeFedConfigs using mutating admission webhook.
- #1038 Removed template validation schema from Federated API's to facilitate upgrade scenarios.
- #690 Extends
kubefedctlby adding theorphaning-deletioncommand, which allows toenableordisableleaving managed resources, when their relevant federated resource is deleted. In addition, the command allows to check currentstatusof the orphaning deletion mode. - #1044 If a target namespace
of
federateand allorphaning-deletioncommands is not specified, use the namespace from the client kubeconfig context.
-
v0.1.0-rc324 Jun 2019 pre-release -
v0.1.0-rc227 May 2019 pre-releaseNothing published for this version
-
v0.1.0-rc122 May 2019 pre-releaseRelease notes
Open source →- #887 Updates KubefedConfig API based on Kubernetes API conventions.
- #885 Updates FederatedTypeConfig API based on Kubernetes API conventions.
- #886 The ca bundle for a member cluster is now stored as an optional field of KubefedCluster since a ca bundle may not be required for all clusters and is not sensitive information that requires storage in a secret.
- #865
github repo is renamed to kubefed. Following changes are done:
- API group changed from federation.k8s.io to kubefed.k8s.io
- FederatedCluster API is renamed to KubefedCluster
- FederationConfig API is renamed to KubefedConfig
- golang imports changed to use vanity url sigs.k8s.io/kubefed
- docker image is renamed to quay.io/kubernetes-multicluster/kubefed
- helm chart is renamed to kubefed
- All role, rolebindings, service-account begining with federation- prefix renamed to kubefed-
- #875 Insecure member clusters are no longer supported due to KubefedCluster.SecretRef being made a required field.
- #869 The api endpoint of a member cluster is now stored in KubefedCluster instead of a cluster registry Cluster.
- #688 Cluster references in placement are now objects instead of strings to ensure extensibility.
- #832
kubefedctl federatecan take input from a file via--filenameoption and stdin via--filename -option. - #868
The default kubefed system namespace has been changed from
federation-systemtokube-federation-system. Thekube-prefix is reserved for system namespaces and including it avoids having the kubefed namespace conflict with a user namespace. - #740 Propagation status is now recorded for all federated resources.
- #844
kubefedctl federatenamespace with its content gets an option to skip API Resources via--skip-api-resources.
-
v0.0.1006 May 2019Release notes
Open source →- #722 - Removal of the FederatedTypeConfig for namespaces now disables all namespaced sync controllers. Additionally, the namespace FederatedTypeConfig must always exist prior to starting any namespaced sync controller.
- #612 - Label managed resources in member clusters and only watch resources so labeled to minimize the memory usage of the federated control plane.
- #721 - kubefed2 disable now deletes the FederatedTypeConfig rather than set propagationEnabled, waits for the sync controller to shut down, and optionally removes the federated type CRD.
- #825 - kubefed2 tool is renamed to kubefedctl.
- #741 -
Added conversion of a namespace and its contents to federated
equivalents via
kubefedctl federate ns <namespace> --contents.
-
v0.0.929 Apr 2019Release notes
Open source →- #776 -
Switch to use
scopeinstead oflimitedScopeto specify if it isNamespacedorClusterscoped federation deployment. - #797 - Cross-cluster service discovery now works for multi-zone clusters. There is an update to KubefedClusters and ServiceDNSRecord API types wherein the zone field is changed to zones.
- #720 -
kubefed2 enablenow succeeds if federation of the type is already enabled. - #738 -
Cleanup
kubefed2 enablerequired arguments and remove unnecessary--registry-namespaceoption fromkubefed2 <enable|disable>. - #737 - Switch to use KubefedConfig resource rather than command line options for kubefed controller configuration management
- #549 - As a result of watching only labled resources, unlabled resources in unselected clusters will no longer be deleted.
- #663
kubefed2 federatenow supports the--enable-typeflag to optionally enable the giventypefor propagation.
Release notes
Open source →- #776 -
Switch to use
scopeinstead oflimitedScopeto specify if it isNamespacedorClusterscoped federation deployment. - #797 - Cross-cluster service discovery now works for multi-zone clusters. There is an update to FederatedClusters and ServiceDNSRecord API types wherein the zone field is changed to zones.
- #720 -
kubefed2 enablenow succeeds if federation of the type is already enabled. - #738 -
Cleanup
kubefed2 enablerequired arguments and remove unnecessary--registry-namespaceoption fromkubefed2 <enable|disable>. - #737 - Switch to use FederationConfig resource rather than command line options for federation controller configuration management
- #549 - As a result of watching only labled resources, unlabled resources in unselected clusters will no longer be deleted.
- #776 -
Switch to use
-
v0.0.810 Apr 2019Release notes
Open source →- #652 - Switch to sourcing the template for a FederatedNamespace from a field rather than the containing namespace. This ensures uniformity in template handling across all federated types.
- #716 -
Upgrade kubebuilder version to v1.0.8
- Removed generated typed clients for federation apis from tree. Use generic client to operate on federation apis as shown here
- Helm based deployment method will be the only supported deployment method to install kubefed control plane.
- #622 -
Switched the sync controller to using a new finalizer
(
kubefed.k8s.io/sync-controllerinstead offederation.kubernetes.io/delete-from-underlying-clusters) and replaced the use of the kubeorphanfinalizer in favor of an annotation to avoid conflicting with the garbage collector. This change in finalizer usage represents a breaking change since resources reconciled by previous versions of the sync controller will have the old finalizer. The old finalizer would need to be manually removed from a resource for that resource to be garbage collected after deletion. - #698 -
Fix the generated CRD schema of scalable resources to define the
retainReplicasof typebooleanrather than the invalid typebool. - #662 Federating a resource could now function without a federation API.
- #661 Accept group qualified type name for type in federate resource.
- #660
kubefed2 federatehas been updated to support output to yaml via-o yaml. YAML output would still require a Kubernetes API endpoint to function.
Release notes
Open source →- #652 - Switch to sourcing the template for a FederatedNamespace from a field rather than the containing namespace. This ensures uniformity in template handling across all federated types.
- #716 -
Upgrade kubebuilder version to v1.0.8
- Removed generated typed clients for federation apis from tree. Use generic client to operate on federation apis as shown here
- Helm based deployment method will be the only supported deployment method to install federation control plane.
- #622 -
Switched the sync controller to using a new finalizer
(
federation.k8s.io/sync-controllerinstead offederation.kubernetes.io/delete-from-underlying-clusters) and replaced the use of the kubeorphanfinalizer in favor of an annotation to avoid conflicting with the garbage collector. This change in finalizer usage represents a breaking change since resources reconciled by previous versions of the sync controller will have the old finalizer. The old finalizer would need to be manually removed from a resource for that resource to be garbage collected after deletion. - #698 -
Fix the generated CRD schema of scalable resources to define the
retainReplicasof typebooleanrather than the invalid typebool.
-
v0.0.722 Mar 2019Nothing published for this version
-
v0.0.621 Feb 2019Nothing published for this version
-
v0.0.515 Feb 2019Nothing published for this version
-
v0.0.410 Dec 2018Nothing published for this version
-
v0.0.313 Nov 2018Nothing published for this version
-
v0.0.219 Oct 2018Nothing published for this version
-
v0.0.2-rc.124 Sep 2018 pre-releaseNothing published for this version
-
v0.0.2-0.20180924154508-5f7c088f301f24 Sep 2018 pre-releaseNothing published for this version
-
v0.0.113 Jul 2018Nothing published for this version
-
v0.0.1-test-fake11 Jul 2018 pre-releaseNothing published for this version
-
v0.0.0-20180711220825-9bcc7247f18811 Jul 2018 pre-releaseNothing published for this version